About
Jesse Read is a Professor (HDR) in the ORAILIX team of the Data Analytics and Machine Learning pôle at École Polytechnique, part of Institut Polytechnique de Paris. He leads research in machine learning with a focus on multi-label classification, data streams, and reinforcement learning, with applications spanning medicine, energy, and transportation systems.
His research interests center around two major themes: multi-label and multi-target prediction (modeling multiple interconnected tasks), and learning from sequential data and data streams. He has particular interest in explainability, uncertainty analysis, robustness, domain shift, transfer learning, and continual learning. His methodological toolkit includes deep neural networks, probabilistic graphical models, Monte Carlo methods, and classical machine learning algorithms.
Professor Read has built a strong research program with significant funding, including an ANR grant for Dynamic Graph Signal Processing in 2024. His work on Classifier Chains for Multi-label Classification received the prestigious Test of Time Award at ECML-PKDD 2019. His research bridges theoretical advances with practical applications across diverse domains.
- Test of Time Award at ECML-PKDD 2019 for Classifier Chains for Multi-label Classification
As an educator, Read coordinates multiple courses including Deep Reinforcement Learning, Reinforcement Learning and Autonomous Agents, and Machine Learning. He co-coordinates the Trustworthy and Responsible AI (TRAI) Masters program and the Artificial Intelligence and Data Science Track. He has supervised numerous PhD students whose work spans predictive maintenance, reinforcement learning for routing problems, rail network modeling, and medical applications. His River library has become a standard tool for stream learning in Python, resulting from the merger of Creme and scikit-multiflow.
Research fields
